I looked in zk under /accumulo/<uuid>/tracers and there were no children there. The error messages continued to happen, so that didn't seem to be the issue. FWIW, I'm seeing the same stacktrace from the master, multiple tservers, and the gc.
On Tue, Feb 25, 2014 at 4:36 PM, Eric Newton <[email protected]> wrote: > Try removing the tracer's entry in zookeeper. > > The spamming should have stopped when the ephemeral node advertising the > tracer's location expired. > > -Eric > > > > On Tue, Feb 25, 2014 at 4:27 PM, Sean Busbey <[email protected] > >wrote: > > > On Tue, Feb 25, 2014 at 2:16 PM, Mike Drob <[email protected]> wrote: > > > > > What is the recommended way to disable tracing on a cluster? I tried > just > > > killing my tracers, but now my tablet servers are spamming the monitor > > logs > > > with ConnectionRefused exceptions. A cleaner method to do this would be > > > nice. > > > > > > > You mean for the internal system tracing rather than just configuring all > > the tables to not do tracing, right? > > > > I thought configuring a user/password that didn't have access gave an > > acceptably small amount of logged feedback, but I haven't tried it in 6-9 > > months. > > >
